Transient Photocurrent Response in a Perovskite Single Crystal‐Based Photodetector: A Case Study on the Role of Electrode Spacing and Bias

Transient photocurrent is a widely applied characterization technique to probe the charge‐carrier photogeneration and extraction dynamics in perovskite optoelectronic devices. Despite the large number of studies on the properties of perovskite single‐crystals (SCs) photodetectors (PDs), the underlying mechanism that governs the spectral line shape of transient photocurrent is not fully understood. Here, methylammonium lead bromide (MAPbBr3)SC based PDs are used to study the effect of different electrode spacing and bias on the transient photocurrent response under blue and green light irradiation. The observed differences in the spectral line shape of the transient photocurrent are explained using three‐step carrier transport model, which reveals the occurrence of carrier trapping and a recombination process in MAPbBr3 SC. The findings are further corroborated by intensity‐dependent photocurrent and impedance spectroscopy analysis of the resulting PDs. This work provides a basic insight into the origin of the different behavior of transient photocurrent response under variable electrode distance, bias, and irradiance light, which is expected to help to further understand and optimize the performance of perovskite based PDs.


Introduction
Photodetectors (PDs) have drawn significant interest from the research community due to their wide range of use in optical www.advancedsciencenews.com www.advelectronicmat.de of defects or trap states and ions accumulation on the charge carrier transports in MHP SC based PDs. [20,[31][32][33][34] Another technique to study the time-dependent transportation and extraction of photogenerated charges in semiconductor based devices is transient photocurrent (TPC). [29,[35][36][37][38][39] In an ideal condition, the transient photoresponse of PD should be a perfectly half-square shape with a very fast response time (Scheme 1a). However, the shape of this spectrum can be distorted due to the mixed electronic-ionic nature of MHP and presence of large density of defects. By analyzing the literature related to the MHP based PDs, we found that there are majorly two types of transient photoresponse spectral line shapes. When the light is switched on, the photocurrent may first slowly rise to the steady-state value and then decay (Scheme 1b). Alternatively, transient photocurrent shows overshooting sharp transient peaks followed by photocurrent decay for a longer time scale (Scheme 1c). The spectral line shape of the transient photocurrent response was found to be affected by changing the light wavelength, electrode spacing, and applied bias. [29,35,40] However, to the best of our knowledge, there is no study that explains the origin of the shape of the transient photocurrent in MHP based PDs.
Here, utilizing methylammonium lead bromide (MAPbBr 3 ) SC based PD, the effect of various factors on the transient photocurrent response is explored. First, we investigate the role of electrode distance on the transient photocurrent of our PDs irradiated with blue and green light. Under blue light irradiation, the transient photocurrent gradually increases until reaching a steady-state and time to reach this state decreases with increasing the electrode spacing. We find that this primary photoresponse arises from the surface layer, which is influenced by the carrier trapping and density of injected charges from bulk layer. In turn, under green light irradiation, the transient photocurrent quickly reaches the steady-state and starts to decay for a longer time scale. It is found that the charge recombination dominates during transient photocurrent measurement, and decay rate of photocurrent increases with increasing the electrode spacing. The light intensity-dependent photocurrent study reveals that the decrease in electrode distance can reduce the charge recombination and increase the photocurrent density of PDs. The observed charge trapping and recombination behavior under different light irradiation is further supported with the help of impedance spectroscopy (IS). Last, we explore the effect of applied bias on the transient photocurrent response under blue and green light and correlate it with the recombination process. We also show that the surface passivation of MAPbBr 3 SC affects the spectral line shape of the transient photocurrent under blue light.

The Effect of Electrode Spacing on the Transient Photocurrent Response
MAPbBr 3 SC was synthesized using an inverse temperature crystallization method. [41] Figure S1, Supporting Information, displays the image of the fabricated SCs and X-ray diffraction (XRD) of the grounded MAPbBr 3 SC. The XRD analysis confirmed the presence of pure single-phase cubic perovskite phase of Pm3m space group with lattice constant a = 5.92 Å, which matches well with the literature. [42,43] A set of planar MAPbBr 3 SC PDs with various Pt electrode spacings was fabricated according to our recently published report (for details, see Experimental Section). [44] First, the transient photoresponse of the resulting PDs was measured at a fixed bias voltage of 2 V and under low illumination intensity (1 mW cm −2 ) of blue light ( = 448 nm). As shown in Figure S2, Supporting Information, the value of photocurrent increases with increasing the electrode spacing due to increase in the active area, which is consistent with the previous reports. [29] From Figure 1a (normalized spectra), it is observed that the spectral line shape of the transient photocurrent changes with increasing the electrode spacing and can be divided into three components: i) fast rise time to reach the maximum photocurrent (zone A); ii) rate of photocurrent loss from the maximum value (zone B); and iii) photocurrent decay after switching off the light (zone C). At the beginning of irradiation, the photocurrent gradually increases until it reaches a steady-state value and then slowly decreases for all PDs. The time to reach a maximum photocurrent is found to equal 325, 175, and 125 ms for the device with 100, 150, and 200 μm electrode spacings, respectively ( Figure S3, Supporting Information). However, one should expect the fastest photocurrent saturation in the device with 100 μm electrode spacing due to its lowest active area (Table S1, Supporting Information).
To explain this behavior, we took into account the three-step carrier transport model, which was recently proposed by Xing et al. [45] In this model, a planar type MAPbBr 3 SC based PD is divided into surface layer, surface-bulk transition layer, and bulk region. Upon excitation of a semiconductor, the electrons are formed in the conduction band while holes remain at the valence band. We noted that a maximum concentration of charge carriers was on the surface of SC as the penetration depth of blue light was ≈200 nm in MHPs. [46] Therefore, the contribution of surface-bulk transition layer and bulk region was negligible under blue light irradiation ( Figure 1b). On the other hand, the evaporation of organic component of MHPs during illumination can lead to the formation of excessive lead bromide, which increases the defect density on the surface of SC. [45,47] When the electric field was applied, the extra charges were injected from the bulk to the surface of SC. [45] These charges could be trapped by the lead based defects, which may act as Lewis acid. [47] In addition to extra charges, the deep-level-donor-like point defects inside the bulk could also migrate toward the surface of SC under applied electric field. [47,48] The formation of deep-level-donor-like point defects and injected charges was highly sensitive to the applied electric field. As seen in Table S1, Supporting Information, the value of electric field intensity increases with decreasing the electrode spacing. We can conclude that the higher density of injected charges and deep-level-donor-like defects on the surface of SC are due to increase in electric field. Therefore, longer time to passivate surface defects and reach the steady-state photocurrent is needed for the device with 100 μm electrode spacing.
After reaching the maximum, the photocurrent starts to decay for a longer time scale (zone B). Lafalce et al. demonstrated that the photocurrent response in MAPbBr 3 SC based PD is dominated by the surface charge effects generated by the accumulation of intrinsic ionic impurities such as Pb 2+ and Br − . [49] In our previous study, we also showed that the ion accumulation near the metal electrode affects the photocurrent response of SC based PD. [33] The accumulated ions influence the interface band bending and increase the Schottky barrier height (SBH) between metal and MAPbBr 3 SC, which could be the reason for the observed photocurrent loss from the maximum value. From the photocurrent decay after switching off the light (zone C), we observed that the device with 100 μm electrode spacing has the slowest decay rate compared to the devices with other electrode spacings. By integrating decay curves, the amount of extracted charges after turning off the light was calculated for all PDs (Table  S1 and Figure S3, Supporting Information) and found to decrease with increasing the electrode spacing. [39] The dependence of extracted charges on the electrode spacing can be correlated with the created trap charges on the surface of SC that slowly detrap after switching off the excitation. However, as was discussed above, highest density of injected charges is trapped in case of device with 100 μm electrode spacing due to its highest electric field. Thus, the high density of trapped charges starts to detrap after switching off the excitation, which explains the highest charge extraction in this type of device.
To study the effect of different wavelength of light on the transient photocurrent, we irradiated our PDs with the green light  [45] Copyright 2020, Nature Publishing Group.
(530 nm) and observed only two major components in the resulting transient photocurrent spectra (Figure 2a; Figure S4, Supporting Information). The first component observed under blue light and associated with the carrier trapping at the defect states was negligible under green light (zone A), which suggests higher carrier trapping under blue light irradiation. [35,46] Under green light irradiation, the formation of excessive lead bromide on the surface of SC was suppressed due to its higher penetration depth (≈400 nm) and reduction in photon energy (Figure 2b). When formed, these defects were cured by the injected charges and no charges were longer trapped after few ms of the irradiation process. Consequently, the surface-bulk transition layer could play a significant role in the transient photocurrent response under green light irradiation, and the probability of charge recombination in the surface and surface-bulk transition layers was enhanced. [50] In turn, the photocurrent loss with time (zone B) under green irradiation could be attributed to the combined effect of charge recombination and ions accumulation due to increase in Schottky barrier height between metal electrode and MAPbBr 3 SC. The ion accumulation depended on the intensity of the electric field and increased with increasing the electric field. [33,51] Thus, we should expect the fastest photocurrent decay for the device with 100 μm electrode spacing. However, the fastest photocurrent decay was observed for the device with 200 μm electrode spacing.
To explain this contradictory behavior, we determined the power exponent (J ph ∝ P , where P is irradiation power and is a recombination-related constant under illumination) from the logarithmic plot of photocurrent density (J ph ) as a function of green light irradiation intensity for PDs with various electrode spacings. [44] As shown in Figure S5a, Supporting Information, the value of the exponent increases with decreasing the electrode spacing, and its higher value indicates lower recombination and better charges collection efficiency, which helps to increase the J ph . [52] A similar trend was also observed for the power exponent extracted from the logarithmic plot of J ph as a function of blue light irradiation intensity ( Figure S5b, Supporting Information). The dependence of charges collection efficiency on the electrode spacing could be correlated with the reduced probability of charge recombination as the effective area of surface-bulk transition layer decreased with decreasing the electrode distance. Thus, the charge recombination dominated during the transient photocurrent behavior in zone B, and the device with 100 μm electrode spacing showed the slowest photocurrent decay due to the lowest rate of charge recombination.
The behavior of photocurrent decay after switching off the green light (zone C) is similar to that decay observed after switching off the blue light. The amount of extracted charges from the studied PDs also decreases with increasing the electrode distance. Table S1, Supporting Information, compares the area of decay curves for all devices after switching off the light for all PDs. It is observed that the area under decay curve after switching off the green light is lower than that area after switching off the blue light for all electrode distances. Therefore, the formation of trapped charges in PD under green light irradiation is suppressed compared to that of trapped charges formed under blue light irradiation. This result also justifies our finding that the role of charge recombination in the surface and surface-bulk transition layer dominates the transient photocurrent response under the green light irradiation. To determine the reproducibility of results, the transient photocurrent is measured for another two individual MAPbBr 3 SC based PDs under 1 mW cm −2 illumination intensity of blue and green light. As shown in Figure  S6, Supporting Information, all devices show similar trend in the spectral line shape, confirming the test results' validity.
Next, alternating-current (AC) impedance spectroscopy (IS) was measured to shed more light on the recombination process under different light irradiations in frequencies ranging from 100 mHz to 1 MHz and at 2 V. Figure 3a shows the EIS spectra of PDs with 200 μm electrode spacing recorded under 5 mW cm −2 intensity of blue and green light (for the impedance spectra of PDs with other electrode spacing, see Figure S7, Supporting Information). The shape of the Nyquist plot appears as a semicircle at the high frequency region followed by a small semicircle at the low frequency region. This type of Nyquist plot is wellestablished for the perovskite-based devices (e.g., solar cells, LED, and PDs), and the arc at the high frequency region can be related to the recombination process. [53,54] The recombination resistance equals to 3000 kΩ and 1300 kΩ for the device under blue and green light, respectively. The lower recombination resistance under green light further supports the above results obtained for the transient photocurrent decay and exponent values. This trend is similar for other electrode distances, which are shown in Figure S7, Supporting Information. Figure 3b compares the capacitance responses of PDs recorded as a function of frequency under blue and green light irradiation. It is well known that the low-frequency capacitance is mostly associated with the accumulation and migration of ions, while the high-frequency capacitance is considered as geometric capacitance. [55,56] At low frequency range (<100 Hz), a higher capacitance is observed for green light, indicating higher charge accumulation. In turn, the high frequency capacitance remains unchanged under both light irradiations, suggesting similar dielectric properties. Interestingly, the values of exponent under green light are lower than those values under blue light for all electrode distances ( Figure  S5, Supporting Information). It can be correlated with the enhanced probability of charge recombination under green light due to higher contribution of surface-bulk transition layer. Figure 4a shows the bias-dependent behavior of the transient photocurrents of the device with 100 μm electrode spacing and under 1 mW cm −2 illumination intensity of blue light (for nonnormalized spectra, see Figure S8, Supporting Information). As seen, the transient photocurrent first shows a gradual increase until it reaches a maximum value (zone A), which is in line with the results shown in Figure 1a. Moreover, time to reach maximum value of photocurrent decreases with increasing the bias, suggesting faster charges trapping at the defect states at higher bias ( Figure S9, Supporting Information). It can be correlated with the formation of high density of injected charges on the surface with increasing the bias. Therefore, the surface defects can be quickly cured by these injected charges under high bias. After reaching the maximum value, the photocurrent starts to decay due to the accumulation of intrinsic ionic impurities (zone B). Under applied high bias, these ionic impurities can move faster and accumulate near the electrode. In zone C, it is observed that the area under photocurrent decay curve increases with increasing the bias (Table S2 and Figure S9, Supporting Information). Therefore, the amount of extracted charges from PDs also increases with increasing the applied bias.

The Effect of Applied Bias on the Transient Photocurrent Response
When the device is irradiated with the green light, the first component of transient photocurrent (zone A) is only visible under 0.5 V bias (Figure 4b; Figure S10, Supporting Information). A small density of injected charges is formed at low bias, which slowly passivates the surface defects. In zone B, the rate of photocurrent decay increases with increasing the applied bias. This can be explained by the concept of high charge accumulation near the electrode under applied high bias, which enhances the ionic contribution in photocurrent. As demonstrated in Section 2.1., the ions accumulation influences the interface band bending and increases SBH between metal and MAPbBr 3 SC. The observed fast photocurrent loss under high bias can be attributed to the increased SBH. Table S2, Supporting Information, compares the area of decay curves after switching off the light under applied different bias, and the device under 2 V bias shows the slowest decay rate (zone C). Therefore, the formation of trapped charges increases with increasing the bias due to the higher density of injected charges.
To further understand the effect of bias on the recombination mechanism in the MAPbBr 3 SC based PD, the photocurrent characteristics of PD at different irradiance powers are measured. Figure S11a,b, Supporting Information, shows the logarithmic plot of the photocurrent density (J ph ) versus irradiation intensity for PD with 100 μm electrode distance as a function of applied bias under blue and green light. As seen, the value of exponent increases with increasing the applied bias. It can be attributed to the higher efficient charge separation and extraction at highapplied bias due to a decrease in recombination and higher density of injected charge, which helps to passivate the defects. As a result, more photogenerated charges are able to move to the electrode, which is the reason for the observed high J ph in PDs under high bias.
Last, the surface passivation of MAPbBr 3 SC was applied to ensure that the surface defects played a role in the spectral line shape under blue light. Phenylethylammonium bromide (PEABr) was selected as passivation agent due to its confirmed role on reducing surface defects in perovskite SC. [57] As shown in Figure S12, Supporting Information, the spectral line shape of the transient photocurrent under blue light charges after passivation, and there is no such difference between the spectral line shape of the transient photocurrent under blue and green light. The first component observed under blue light, which is associated with the carrier trapping at the defect states (zone A), is negligible after the passivation due to the reduction of the defect states on the surface. These results confirm that the surface defects play a role in the spectral line shape of the transient photocurrent under blue light.

Conclusion
In summary, we investigated the effect of different electrode distances and applied biases on the transient photocurrent response of planar MAPbBr 3 SC based PD. The spectral line shape of the transient photocurrent was divided into three components, which allowed for distinguishing the difference in the charge transport process under blue and green light irradiation. Under blue light irradiation, the initial slow increase in the photocurrent with time was attributed to the carrier trapping at the defect states, which led to surface passivation. The time to achieve a steady state photocurrent decreases with increasing the electrode distance due to the faster carrier trapping at the defect states under a higher electric field. After turning off the blue light, the www.advancedsciencenews.com www.advelectronicmat.de amount of extracted charges decreased with increasing the electrode spacing. On the other hand, the effect of carrier trapping was negligible under the green light, and the transient photocurrent was dominated by the charge recombination and ions accumulation near the electrode. It was observed that the decay rate of photocurrent of PD increased with increasing the electrode distance. Thus, the device with the lowest electrode spacing (100 μm) showed the slowest photocurrent decay and highest photocurrent density due to the reduced charge recombination rate. By comparing IS under blue and green light irradiation, we obtained a picture of the changes in capacitance that indicated the higher probability of charge recombination under green light due to higher contribution of surface-bulk transition layer. In addition, we studied the effect of applied bias (from 0.5 to 2 V) on the transient photocurrent response of MAPbBr 3 SC based PD under blue and green light. While a fast carrier trapping at the defect states was observed under applied high bias and blue light, the carrier trapping process under the green light and high bias was negligible. The rate of photocurrent decay under the green light increased with increasing the applied bias due to faster intrinsic ions migration and accumulation near the electrodes, which enhanced the photocurrent loss with time. In addition, the initial slow increase in the photocurrent with time under blue light, which was attributed to the carrier trapping at the defect states, became negligible after passivation of the surface of perovskite SC with PEABr. This study provides a basic insight into the origin of the different behavior of transient photocurrent under variable electrode distance, bias, and irradiance light, which will help to further understand and optimize the performance of perovskitebased PDs.

Experimental Section
Synthesis of MAPbBr 3 SC: 1.2 m MAPbBr 3 solution was prepared by dissolving 1.32 g of PbBr 2 (≥98%, Sigma-Aldrich) and 0.40 g of MABr (>99%, Sigma-Aldrich) in 3 mL anhydrous DMF for 24 h at room temperature. Next, the clear solution was filtered using a 0.22 μm PTFE filter and sealed in a bottle for 4 h at a constant 80°C temperature. The formed single crystals were collected and washed by isopropyl alcohol, dried at 100°C for 15 min, and stored in a humid-free chamber. Surface Passivation of MAPbBr3: The surface passivation with PEABr was performed according to the recent procedure developed by Chen et al. [57] Specifically, the 10 mL passivation solution was prepared by dissolving 10 mg PEABr into 10 mL tert-amyl alcohol and filtering using 0.2 μm pore size PTFE filter. For surface passivation, MAPbBr 3 SC was soaked for 10 min in 2 mL of passivation solution. Next, the crystal was slowly stirred in 2 mL cyclohexane for another 10 min to remove additional residues and dried at 80°C for 10 min.
Fabrication and Characterizations of the SC-Based PDs: The planar PDs were fabricated based on perovskite SCs by depositing 100 nm-thick platinum (Pt) with variable spacing (100, 150, and 200 μm) by a sputtered magnetron (Leica EM MED020) on (100) facets of the SC. Figure S13, Supporting Information, shows the optical image and schematic representation of the fabricated MAPbBr 3 SC based PD. The total sensing areas of the detectors with variable electrode spacing are summarized in Table S1, Supporting Information. All dark current-voltage (I-V) characteristics, response, and impedance spectroscopy (IS) measurements were performed on a LASC probe station connected to a Bio-Logic SP-150e potentiostat. The illumination power and chopper speed of blue ( = 448 nm) and green ( = 530 nm) LED (Luxeonstar) were controlled by the 2nd channel of Bio-Logic SP-150e potentiostat, which was optimized with the help of a spectrometer (Thorlabs GmbH., PM 100D).

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.